28^th of March

Bhabsadh finally finished his timeout-laden journey. I can focus on the 6 remaining games (well, 5, as I’m a rook up against nemer).

White just played Re1. My plan yesterday was Re8 to prepare king’s escape to f8, but this now needs more calculation.

Things seem to add up.

Later in the morning, another point is in the bag:

This could have gone better.

Also, not a game where I had to work too hard. The engine is also a bit harsh with its evaluation as I prioritized safety and brain energy saving over speed of conversion.

5 games left. The two games against saveliy look promising – pawn up in the white game, and piece up (but defending) in the black game. 

Another result in the group. Saveliy blundered and then timed out the game against the leader: disappointing. It increasingly looks like I’ll need to score at last 1.5/2 in my two games against pruett.

Back to the games.

My original plan here was to play Kf6, and look for creative ways to hold on to the pawn (which is likely not doable). However, what about Kh4? The move looks awkward and dangerous, but I don’t see a way for white to keep my king locked on the h-file, and make progress elsewhere. Time to analyze.

The idea is: hold the bishop on c5 and move the king between h3 and h4. If white threatens mate from h1, move the king to g5. This means that in order to break, white has to move the king away to d5 or a8 to do the rook for bishop + pawn exchange.

However, there I see a win for white unfortunately. Advance the pawn to a6, move the king all the way to a8, and then promote one move ahead of the black h-pawn. If black starts goosestepping the king and the pawn down the h-file, come back with the king. As soon as the king gets to h3, white king gets to f3, and all kind of mates start popping up.

All in all, Kh4 almost works, but almost does not count. I’ll need to analyze Kf6 later today.

After a long and nervous drive (rain, wind) from SF, I’m back at the game. Coming back to Kh4, I’m not sure if white will find the idea of drawing the black king from h4 by going with white king to a8 and then coming back. But then, a basic rule is to never assume what the opponent sees and does not see.

Now, I have an even more creative idea. What about Kh6? White can only win the h5 pawn by putting the king on h4, which I can always dislodge with Bf2+. The rook can no longer both cut the g-file and attack a7 at the same time. If the white king goes to a8, now I’ll be able to push the pawn before the king. For a standing position, the black king can move from h7 to h6, I just need to make sure that it’s on h6 if the white king is close to h4. In that case, the bishop will be moving on the diagonal. If the rook comes out to d2 for example to handcuff the bishop, the king will go to g5.

Maybe I’m missing something, but it’s late enough, so Kh6 it is.

This is the move I'm most proud about in the entire tournament. The engine actually shows Kh4 as the best option, and only after going 10 moves down the algorithm above does it show black as losing. It was an intricate decision point in an interesting endgame.

Coming back to the main game:

Yesterday, I was looking at g4, but it is very committal. It also doesn’t quite work:

Be4 looks better. It prepares g4 in a subtle way, and limits black’s options. It may provoke f5, but I think the pawn on f5 is easier to pile upon compared to f6.

The engine considers all these options equivalent, and consistently rates all at about +0.4. My understanding is that double pawns are just not that important, and the game bears it out.

29^th of March

A bit of development in the table last night. Saveliy timed out in a bunch of games (though, obviously, not mine).

The last move in the game below was, predictably, Ne5.

Originally, I planned Nxe5, but on closer inspection, I’ve got more options: b5 (to gain space), and Rc8 (to contest the c-file). For the last two, I need to check whether g4 is going to be a problem.

First, some analysis of Nxe5, and why I’m looking for alternatives.

This was exactly correct. This side line takes black from -0.2 to +0.8.

Rc8 has some interesting counterplay:

Overall, I don’t think it’s in my interest to help white get a pawn on e5, so I’ll go with Rc8. b5 can be played later – if white allows.

While the lines are far from precise, the decision was spot on. I wish only the bishop stayed at d6; this is the point I became convinced that Bc5 was a mistake.

Moving to the white game:

The plan here was to pile up on the f5 and h-pawns with Bd3 and Rf4.

Calculations show that I’m likely to end up in the pawn ending, which I’m no longer sure is won for white.

Another option is to play it slow, and swing the bishop to h3 to attack f5, while keeping control over the d-file:

I’m actually leaning towards that. Pawns don’t go backwards, and the fact that the pawn is already on f5 is helpful.

The lines are far from precise, but the decision was sound. The first option led to a draw, while this one kept the game alive, and gave black more options to go astray.

Back to the black game:

White has gone all aggressive on me with g4. It looks a bit scary, but now at least I know it’s unlikely to end up in a draw. The main question is what to take with. Originally, I planned Nxd7, in order to have counterplay with Qh4. However, if white plans on continuing to go all guns blazing with g5, Qxg5 will be risky.

While I don’t see an immediate win for white, there is a massive compensation for the pawn.  If I get no choice, I’d go for this line, but then I can also just play 17...b5 for example.

The comp does not believe in mirages, but rather believes in extra pawns, advocating to chomp the g5 pawn. The line above has massive hole in it, with 19...Be4+, which the engine also points out. 

While I don’t see an immediate win for white, there is a massive compensation for the pawn.  If I get no choice, I’d go for this line, but then I can also just play 17...b5 for example.

However, isn’t it just better to take with the queen?

It looks a bit inferior to the other lines. White did not sacrifice anything here, but has most of the attacking options.

Surprisingly, I calculated something very close to the top line, apart from the last move - Bb5, which makes no sense. The eval is also off - not sure why I thought white is in the driving seat here.

There’s of course also the question of what if white _doesn’t_ play g5? For example, Rg1 to prepare to take on g5 with the rook.

Nxd7 is better in that case too, as I’ll at least have the option of Be7.

Continued here: https://www.chess.com/blog/romank66/chess-com-2023-2nd-round-s01e09-time-for-regrets